Different perf. results for same strategy on MC 4 and MC 5

Questions about MultiCharts and user contributed studies.
vandevenne steven
Posts: 11
Joined: 28 Aug 2007

Different perf. results for same strategy on MC 4 and MC 5

Postby vandevenne steven » 01 Mar 2009

There is a big difference between backtesting results I get for a simple strategy on MC v4.0 and MC v5 beta3.

The strategy tested is very simple:

Buy next bar if the current bar is a downbar
Sellshort next bar if the current bar is an upbar

Teh performance results differ greatly. On v4.0 the strategy comes out succesful with almost any combination of stoploss and profit target.

On v5.0 beta 3 the strategy never turns out profitable.

I have tested with tiny time intervals so the difference cannot be attributed to intrabar tick differences.
I have also used quotes on the same time window.

Moreover the strategy is so simple and generates so many trades that the average results do not differ when testing with different quote series on either MC 4 or MC 5 by itself.

Does someone have any idea what may be causing this.

Thanks,

Steven
[/list]

SUPER
Posts: 646
Joined: 03 Mar 2007
Has thanked: 106 times
Been thanked: 84 times

Postby SUPER » 01 Mar 2009

If you can post your test code then it may become a bit easire to look into it.

Regards
Super

vandevenne steven
Posts: 11
Joined: 28 Aug 2007

Postby vandevenne steven » 01 Mar 2009

Hi Super,

No problem here you go:



inputs: variatie (.2) ;

condition1 = C < O ; // Currentbar bar is red

if time > 0000 and time < 2350 then
begin

if marketposition = 0 then
begin

if condition1 then
Buy ( "Barchg LE" ) next bar at close + variatie stop ;
end;

end;


I am testing this on the April gold contract (GCJ9), 233 tick bars.

Thanks for your reply!

Best regards,

Steven

brodnicki steven
Posts: 407
Joined: 01 Jan 2008
Been thanked: 3 times

Postby brodnicki steven » 01 Mar 2009

Since you're using tick bars make sure barsback setting is the same, I would NOT use "auto" for bars back. Use 50 bars or something,with identical start dates and times .
Have you tried it on minute bars, with identical start dates and times ?

User avatar
Marina Pashkova
Posts: 2758
Joined: 27 Jul 2007

Postby Marina Pashkova » 11 Mar 2009

Hi Super,

No problem here you go:



inputs: variatie (.2) ;

condition1 = C < O ; // Currentbar bar is red

if time > 0000 and time < 2350 then
begin

if marketposition = 0 then
begin

if condition1 then
Buy ( "Barchg LE" ) next bar at close + variatie stop ;
end;

end;


I am testing this on the April gold contract (GCJ9), 233 tick bars.

Thanks for your reply!

Best regards,

Steven
Hi Steven,

brodnicki steven is right. It is important to make sure that you have the same starting point for the calculations (same max bars back). Also, when comparing the performance of the strategy in the two versions of MultiCharts, please make sure that:

1. The symbol has identical settings and sessions in both versions.
2. Data requests are the same (better to use the From...To request to make sure the starting point is the same).
3. Signals should be identical.
4. Strategy settings should be the same.

We could run tests for you is you provided us with the following information:

a) The workspace
b) Complete code (what you posted only has entry conditions
c) Settings for this symbol in QM
d) Screen shots illustrating the problem would be helpful as well.

I look forward to hearing from you.

Thank you.


Return to “MultiCharts”